Log in

No account? Create an account

Previous Entry | Next Entry

Table Selection Part 2

For those that haven't read my 1st post on table selection, it is a must read for any SNG player and can be found here: http://jhub3000.livejournal.com/84368.html

I've been meaning to expand on this for a while but just haven't felt like typing up a long entry. I was also skeptical about sharing how much fish can increase your ROI because I feel like people will abuse this info & chalk up a lot of losing players as "fish" when they really aren't. However, if people realize how much a breakeven player can hurt their ROI this should balance out.

First off, let's summarize the main stats of the 1st table selection article. Each type of player drops your ROI x % for 6 mans:

good player: 4.5%
decent player: 2.8%
barely above breakeven player: 2%

for 9 mans:
good player: 2.8%
decent player: 1.8%
barely above breakeven player: 1.2%

Now I'll go into more detail not only on how the #s change if there's 2 or more of each player, but also add more player types like breakeven, -5%, -10%, etc...

The following #s will be 6 max only. However, if you play 9 mans you can kind of get an idea based on the #s above. Again, each player(s) drops your BASE ROI x %:

Cost of 1 good player: 4.5%
Cost of 2 good players: 8.7%
Cost of 3 good players: 12.57%
Cost of 4 good players: 16.16%
Cost of 5 good players: 19.5%

Obviously if you want to win money you should rarely be playing a 6 max game w/ even 2 good players but those are the #s anyhow.

Cost of 1 decent player: 2.8%
2: 5.45%
3: 8.11%
4: 10.56%
5: 12.9%

Cost of 1 barely above breakeven player: 2%
2: 3.9%
3: 5.87%
4: 7.7%
5: 9.46%

Cost of 1 breakeven player: 1.45%
2: 2.86%
3: 4.24%
4: 5.58%
5: 6.89%

Cost of 1 -5% player: 0.29%
2: 0.58%
3: 0.86%
4: 1.15%
5: 1.43%

Cost of 1 -10% player: +0.87%
2: +1.75%
3: +2.64%
4: +3.55%
5: +4.48%

Cost of 1 -15% player: +2.02%
2: +4.12%
3: +6.3%
4: +8.56%
5: +10.9%

Cost of 1 -20% (definition of a fish): +3.18%
2: +6.54%
3: +10.11%
4: +13.89%
5: +17.91%

Cost of 1 -30% player: +5.49%

Cost of 1 -40% player: +7.79%

In my 1st table selection post I recommended that good players use a base ROI (starting ROI of which to add & subtract these #s from) of 8.56. However, now that there's more #s to consider here, I'd say somewhere around 10 to 11% sounds about right.

It is very clear that since my 1st table selection post the games have gotten much tougher. Therefore, I think the ROIs we use to characterize players need to be updated.

If a player has the following ROI or higher at each level you should chracterize them as a good player when table selecting:

$6s: 11%
$16s: 8%
$27s: 6%
$38s: 5%
$60s: 4%
$114s: 3%
$225s: 2%
$335s: 1%

However, more needs to be considered. First off, when I list these ROIs I'm talking about their ROI in 5-10 seated NL hold em turbo SNGs. An advanced search would need to be done to obtain this info. For regular speed SNGs you can probably bump the ROIs up a bit & for super turbos you can lower them.

I recommend doing year 09 advanced searches on players which you are unsure of. If 1 player has played 3,000 games w/ an ROI of 4% at the $60s spread over the years & 1 player has played 3,000 games w/ an ROI of 4% at the $60s in 2009, I will give much more credit to the latter since the games have gotten progressively tougher over the years.

Then you need to factor in variance. If you read my 1st table selection post then you should have a pretty good idea of just how crazy variance can be. To summarize some stats:

After 500 games there is a 50% chance that you are within 5% of your true ROI
After 1,000 games there is a 67% chance that you are within 5% of your true ROI
After 3,000 games there is a 90% chance that you are within 5% of your true ROI

That means that if a player has an ROI of 4% over 3,000 games, all that really tells you is there is a 90% chance that his true ROI is between 9% & -1%. With sustainable ROIs becoming much smaller (especially at the high stakes) this can make it very tough to characterize players into the right ROI categories. My advice would be to usually just give them the benefit of the doubt. This is 1 thing I've been doing more lately & I think it has helped tremendously. There's plenty of players who have glaring leaks but do other things well enough to make up for these leaks. Downgrading a player too much just because you notice some leaks in their game can be a very costly mistake.

You always need to figure out how ROIs translate at different levels. If you're playing a $119 & a player has played 500 games w/ an average stake of $25 w/ a -10% ROI, you can be pretty confident that he is -20% or more in a $119. Use your head in spots like this.

Now I'll go through some real life examples using sharkscope tournament selector:
Game 1: $119 on FTP:
Username Games Played Av. Profit Av. Stake Av. ROI Total Profit Form Ability /100
Player 1 30 $78 $64 44% $2,337 - 92 x
Player 2 899 -$1 $87 7% -$459 - 73 x
Player 3 6,209 $2 $66 7% $10,531 Tilt 85 x
Player 4 3,573 $2 $52 5% $7,615 - 85

Player 1 has only played 30 games. That doesn't really tell me anything so I'd just go with the default & mark him as a breakeven player. With player 2 you want to pay more attention to his avg. ROi than his total profit. The fact that he is losing overall w/ that ROI just means that he probably lost over a small sample of high stakes games. I advanced search player 2 & see that he's 2% over 771 6-10 seated Nl holdem & -1% over 583 6-10 n 2009. Given these stats combined I would probably characterize him as a barely above breakeven player. Player 3 has not played any 6-10 seated games before 2009. In 6-10 seated he is 4% over 3,855 games. Given that knowledge I would characterize him as a decent player at the very minimum, and probably give him the benefit of the doubt & chalk him up as a good player until he shows me otherwise. With this particular player I know that he is living with & being coached by a top SNG player so I will definitely characterize him as a good player. Player 4 is 7% over 3,237 6-10 seated & 2% over 1,889 in 2009 where his avg. stake has been a tad higher at $70. Again, at minimum I would characterize this guy as a decent player & probably give him the nod as a good player until he proves otherwise.

Now let's calculate our ROI. We have 2 good players so they drop our ROI 8.7%. We have 1 barely above breakeven player so he drops our ROI 2%. We have 1 breakeven player so he drops our ROI 1.45%. If our base ROI is 11% then our ROI in this game is -1.15%. If we wanted the game to be +EV pre rakeback we would want to watch the lobby & see if the 5th player who registers is -12% or worse. If there's better options out there for games or we don't need many more tables we would probably want to be more selective than settling for a breakeven pre rakeback game.

As is with the 4 players currently in the game we could calculate our profit in that game as follows: -0.0115 x $119 = -$1.37. Rake is $9. If the game is on FTP & we have 27% rakeback then rakeback = $2.43. $2.43 - $1.37 = $1.06 profit. That's not exactly the type of profit you should be looking for in a $119. You could almost always drop down & play a lower buy-in game for more profit w/ less variance & stress.

We wait 3 minutes & the following $119 is registering:

Username Games Played Av. Profit Av. Stake Av. ROI Total Profit Form Ability /100
Player 1 1,642 -$4 $53 1% -$6,872 - 69 x
Player 2 2,245 -$2 $47 -4% -$5,204 - 68 x
Player 3 2,772 -$37 $194 -18% -$102,084

Player 1 hasn't played any 6-10 seated pre 2009 & has played 1,158 at 2% in 2009. We will characterize him as a decent player. Player 2 hasn't played any 6-10 seated pre 2009 & is -3% over 1,235 in 2009. I would usually give him the benefit of the doubt & characterize him as a breakeven player until I get reads on him. Player 3 has played 1,467 6-10 seated at -18% & 1,163 in 2009 at -15%. We'll characterize him as a -15% player.

So far our estimated ROI in this game starting w/ a base ROI of 11% is 8.77%. Odds are the last 2 players will be breakeven players, in which case our ROI would be 5.98%. Our total profit for that game would be $9.55, much higher than the 1st game. Even if the last 2 players to register were good players, which is pretty unlikely given the odds, this game would still be more profitable than the 1st game.

Obviously you won't be able to do all of theese calculations in game when multi tabling. Get to know the #s so you have them memorized. Then get to know the players. 1 thing I have found very useful is to look up players in a bunch of games once per day while I'm not playing. I'll then take notes on these players. It shouldn't take long. This is especially useful if you play on full tilt where you can color code the players. Use your head & check your ego at the door.

Since dropping my tables from 10-12 to 8-10 & changing a few things up I am still crushing.

Overall stats:
Games Played Av. Profit Av. Stake Av. ROI Total Profit Form Ability /100 Network Filter
1,787 $9 $116 8% $16,186 N/A N/A Global Alias 11/11/2009 12/4/2009 E6-6 SNG Only

Turbos stats (excluding super turbos):
1,236 $11 $138 10% $14,172 N/A N/A Global Alias 11/11/2009 12/4/2009 E6-6 Spd=T SNG Only

Hourly rate: $244.14

I'm very confident in my game right now.

Year 2009:
SNG Profit: $(13,090.29)
Rakeback/Bonuses: $205,973.63
Total SNG Profit: $192,883.34
Staking: $42,219.80
Total Profit: $235,103.14

Career Sharkscope Stats:
Normal speed 5-6 seated:
Games Played Av. Profit Av. Stake Av. ROI Total Profit Form Ability /100 Network Filter
10,124 $9 $107 12% $89,072 N/A N/A Global Alias E5-6 Spd=N SNG Only
Turbo 5-6 seated:

Turbo 5-6 seated:
66,450 $5 $156 5% $362,188 N/A N/A Global Alias E5-6 Spd=T SNG Only

Super turbo 5-6 seated:
7,768 -$3 $107 -1% -$22,044 N/A N/A Global Alias E5-6 Spd=ST SNG Only

9-10 seated:
1,534 $1 $91 0% $2,246 N/A N/A Global Alias E9-10 SNG Only

12 or more players:
135 $20 $23 101% $2,650 N/A N/A Global Alias E>=12 SNG Only

All games:
86,432 $5 $144 5% $435,648 N/A 89 Global Alias SNG Only

Blogs2Go.net - FREE BLOG Listings

My Zimbio
Top Stories

PlaY Fantasy Poker


( 26 comments — Leave a comment )
Dec. 4th, 2009 08:23 pm (UTC)
Just curious...i trust your judgement with these stats etc etc, but, if everyone followed these table selection rules would the games even fill enough? Slightly above break even players who make there money off rakeback by playing masses of volume surely wouldnt be able to anymore? Also, are you going with the assumption that there is still thousands and thousands of players (obviously 90% fish) out there who don't even know sharkscope exists?
Dec. 4th, 2009 09:48 pm (UTC)
The thing is everyone won't follow these rules. The people who are likely to follow these rules are people who are smart enough to be breakeven or better, which are the same players that hurt our ROI. If fish were smart enough to follow rules like these they wouldn't be playing in the 1st place & would spend time studying the game before playing. There's so much info out there about the game that if they were the type of player who would be proactive about making money at the game (ex. table selecting) they wouldn't be losing in the 1st place.

I also think a good amount of the true fish are just rich people who play for fun & don't care that they're losing because it's insignificant to them.


Edited at 2009-12-04 09:57 pm (UTC)
Dec. 16th, 2009 05:46 pm (UTC)
That's why I also don't see the point in good players blocking their sharkscope stats. If they're out there its only the good players you don't want in your games anyway that are dissuaded. Of course I'm sure some people just want their privacy but that's a different matter. Great post anyway

Dec. 16th, 2009 06:22 pm (UTC)
I completely agree man.

GL at the tables,
Dec. 4th, 2009 10:24 pm (UTC)
Is there a better chance of your timberwolves winning the nba finals this year of kenny05 ever following this rules?
Dec. 4th, 2009 10:32 pm (UTC)
definitely a better chance of Kenny following the rules...he'll learn his lesson...he claims he wants to table select well in 2010...if he goes for elite again I'm not sure that his pokerstars table selection will ever will be that good

Dec. 6th, 2009 02:39 am (UTC)
Nice work as always Hub. The thing that many people don't seem to be able to grasp is that breakeven players aren't good to have at the table.

Also, Stars now has color coded notes if you haven't noticed. It's very entertaining for me to watch games go off with 6 out of 9 guys labeled red.

Dec. 6th, 2009 07:00 pm (UTC)
Thank you. Yeah in all honesty how breakeven players effect your ROI is the thing about table selection I learned last so I was somewhat guilty of it as well, although not as much as others because I knew they weren't good for my ROI, I just didn't know to what degree. I think that's 1 of the things that hurt me this year.

I didn't know that stars has color coded notes. I think every site should have that. It's amazingly helpful. Cake has them as well except the color code doesn't show up in the lobby, only in the game. That doesn't hurt me too much though because I use very little table selection on Cake because of my rakeback deal anyway & I basically know who the top guys are.

Dec. 7th, 2009 12:46 am (UTC)
Any advice on precisely how to use the color coding?
Dec. 7th, 2009 02:26 pm (UTC)
Here is how I do it:
orange = top player
yellow = good/decent player
purple = barely above breakeven player
red = keep an eye on, usually someone winning over a small sample
blue = breakeven or insignificant sample
green = losing player
dark green = fish (-20% or worse at $119)

You could go into a little more detail & breakup the losing players into more groups. I've considered that but I think this is working fine for me.

Dec. 8th, 2009 04:39 pm (UTC)
Questions from a bit of a beginner

I had some very general questions that will probably sound very beginner to most. If anyone has any feedback please respond in the thread or drop a comment on my blog:


Keep in mind I have always been more of a cash game player, but here it goes:

1) Do you use Sharkscope to get the ROI% information on all of your opponents? Is that the best software out there to use in your experience?

2) I thought sharkscope was banned on PokerStars. Has anyone received a warning or even an account freeze?

3) How does the discussed information relate to turbo MTT SNGs? Or are these types of tournaments generally avoided by most profitable SNG players due to the variance involved? I imagine the variance is ridiculous with regards to the MTT SNGs, but I would assume the players you face are MUCH WORSE. Then again it would be very hard and maybe impractical to table select on a 45-man or 180-man SNG.

Thanks for the great post.

Jun. 26th, 2010 03:44 am (UTC)
Re: Questions from a bit of a beginner
It's my first time to post a reply, thanks for your sharing.
Jun. 26th, 2010 08:14 am (UTC)
Re: Questions from a bit of a beginner
I hope you don't mind me linking you so that this information spreads more.
Dec. 9th, 2009 10:38 am (UTC)
nice one
Hey man, really like you blog.

Anyway just a fast question, in a fullring game with 9ppl..
Whats like "ok" to play with i mean by the calculations?
(lowstakes) but still makes it profitable to tableselect i know that for sure. So.. rather then jumping into a game with 6-7regs w8 some and ill have 1-2. But what about if like 3 is registered lets say they all have 5% roi? how much does it affect my roi? or what about if 4 has and theres 2x 5% 1x 9% and 1 breakeven reg.
Ill read the post again and try myself aswell. Thanks alot!
Dec. 9th, 2009 06:22 pm (UTC)
Help getting money on Cake
Hey man great blog really do enjoy reading. Quick question, I know it always says in the beginning of your blog that we can ask you if we need help getting money onto cake, I was wondering if I could email you to have you help me out with this. Just wondering what your email is so I can talk to you about this later. Thanks and keep up the good work
Dec. 10th, 2009 03:02 pm (UTC)
Re: Help getting money on Cake

Dec. 9th, 2009 11:51 pm (UTC)
good entry hub, ty
Jun. 26th, 2010 02:43 pm (UTC)
Definitely recommending this all around.
Dec. 10th, 2009 12:47 am (UTC)
Hi buddy, i've already signup at cake poker with code jhub, my screen name is zachen1989. And i need help getting money into it, so i can start with SSS. thanks a lot , my screen name is zachen1989 !!
Dec. 10th, 2009 03:03 pm (UTC)
email me: jhub3000@hotmail.com

Dec. 21st, 2009 08:51 am (UTC)
Tremendous post! I am glad to see you have been crushing as of late and hope the new year will bring you much joy and luck!
Dec. 24th, 2009 04:34 pm (UTC)
can you put a mouse link again old not working .thanks.
Dec. 28th, 2009 02:49 pm (UTC)
Re: J

They've changed this model a little since I got it but it looks like it's still pretty similar.

Dec. 28th, 2009 07:04 am (UTC)
Great post
You have the best blog jhub. Very informative, I have been table selecting ever since your first post on table selection. Thanks for your insight!
Dec. 22nd, 2010 01:44 pm (UTC)
yo its comma d iv had a great year and i would attribute it to table selection more than anything. but with these #s most games r unbeatable. theres no way u could get 10 games goin with those standards. is ftp that much softer? and how do u explain all the high stakes grinders who dont table select at all still winning? i cant understand how some people r winning at the highest $ w/o even lookin who they r playin against. i see 500s w/ 5 top players registered all the time and these guys r playin tables like this over and over and still winning, not to mention all the regs who dont tab select anything and play too many tables who r playin far less than optimaal who still win over rb. do u just atribute this to variance ? and do u really think guys w/ small samples r actually break even? that would make it damn near impossible to get 10 good tables ata time...
Dec. 22nd, 2010 08:40 pm (UTC)
Re: comma_d

I'm glad you're having a great year. You're def 1 guy I root for.

As for the #s, remember these are for super turbos, whereas you play turbos & regular speeds. There's more edge in turbos & normal speed so the #s are different. In STs you start w/ 10BBs. There's not a whole lot of edge there, especially in this day and age when some of the fish even have a clue w/ those stacks. Edit: I just realized you replied to my old table selection post on turbos rather than my new one on STs but whatever. It all still applies for the most part.

I not only get 10 games going, but I get 16 games going w/ strict standards like this. I just have to play a wide range of stakes & my avg. buy-in is less than $50 lately. I would probably be willing to bet anyone that I'm top 5 in 6-10 seated profit next year. As for FTP being much softer, I doubt it. The high games on FTP are regularely stocked w/ the same 5 players, just like the ones you speak of on stars.

As for the high stakes grinders winning w/ no table selection, it's variance. Let's take a look at some #s:

Here's the $301-$1,000 stats of the top 10 from the 5-6 seated $301-$1,000 ST leaderboard:
ftp301to1000stlb 49,436 $6 $179 2% $280,714 N/A N/A Global Alias E5-6 G=H NoLim Spd=ST NormalFormat

Here's the stats of the same exact players under $300:
FTP301to1000STlb 40,536 $1 $123 0% $22,810 N/A N/A Global Alias

They're 2% at $300+ & 0% at $0-299. Enjoy it while it lasts.

Here's the $300+ stats of the 5-6 seated ST $101-300 LB:
FTP101to300STlb 8,313 -$9 $447 -3% -$73,230 N/A N/A Global Alias

The $101-300 leaderboard is more likely to tell you who the best players are because they're putting in larger samples. Unfortunately for them they don't have the run good power of the other players.

There's guys like leoc00 & Kenny05 that you see towards the top of the high stakes LBs.

Here's leoc00's lifetime $300+ 5-6 seated stats:
leoc00 6,470 -$3 $502 0% -$20,410 N/A N/A Global Alias E5-6 S300-NaN G=H NoLim NormalFormat

Search 2009 & beyond & he's -1% despite being on the $301-$1,000 5-6 seated ST leaderboard this year.

Here's Kenny's $300+ 5-6 seated stats this year:
5,069 $11 $475 2% $56,954

Here's his under $300 5-6 seated stats this year:
3,164 -$2 $205 -1% -$7,476

I think you get the point.

The reality is that you need 25,000 games just to be sure you're within 2% of your true ROI. By the time anyone actually gets that big of a sample at the high stakes the games will have changed drastically. For example, leoc00 has not more than 6,000 $300+ 6 max lifetime & he's 1 of those guys that's blindly registering for every one of them.

As far as thinking players w/ small samples are actually breakeven, well it's all in the math. If they have more than 500 games they're likely to be within 5% of their true ROI. If a player is 5% he's not likely to be less than 0% long run. If he's 2% than he's not likely to be less than -3% long run. A -3% player is still going to drop your base ROI 1.11% in a ST. Remember, base ROI is your ROI BEFORE you add or subtract any percentage points for your opponents. In reality these guys w/ small samples may be regs on other sites just mixing in some games or players w/ a lot of leaks but only play 2 tables at a time & are able to make up for it. Even if a guy isn't playing like a reg I don't think you can just write them off. There's definitely some unconventional players who have won over big samples.


Edited at 2010-12-22 08:46 pm (UTC)
( 26 comments — Leave a comment )